The AI Museum
AI Analytics1,072 recordsRecounted 23 September 2026

Open-weight language models from China

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
3 filters active, showing 63 of 1,072 recordsClear all

63

Models

16 organisations

100%

Open weights

63 open, 0 closed

0

Frontier

By Epoch's flag

2.4T

Largest traceable

Qwen3.8-2.4T-A95B

2020–2026

Span

39 with a traceable size

Releases

by year
2020202120222023202420252026

Organisation

Country

Domain

Records

newest 50 of 63
Open-weight language models from China, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
DeepSeek V4.1 FlashDeepSeek2026-09748B *Open
GLM-5.3-FlashZ.ai (Zhipu AI)2026-08320B *Open
Qwen3.8-2.4T-A95BAlibaba2026-082.4TOpen
Tencent Hy4 previewTencent2026-08770B *Open
Kimi K3Moonshot2026-072.8T *Open
Tencent Hy3Tencent2026-07295BOpen
LongCat-2.0Meituan Inc2026-061.6T *Open
MiniMax-M3MiniMax2026-06428BOpen
DeepSeek-V4-FlashDeepSeek2026-04284B *Open
DeepSeek-V4-ProDeepSeek2026-041.6T *Open
GLM-5.1Z.ai (Zhipu AI)2026-04754BOpen
Kimi K2.6Moonshot2026-041.0TOpen
MiMo-V2.5-ProXiaomi Corp2026-041.0T *Open
Tencent Hy3 previewTencent2026-04295BOpen
GLM-5Z.ai (Zhipu AI)2026-02744B *Open
Kimi K2.5Moonshot2026-021.0T *Open
Qwen3-Coder-NextAlibaba2026-0280B *Open
Qwen3.5 397B-A17BAlibaba2026-02397B *Open
Qwen3.5-122B-A10BAlibaba2026-02122B *Open
GLM-4.7Z.ai (Zhipu AI)2025-12358B *Open
MiniMax-M2.1MiniMax2025-12229BOpen
DeepSeekMath-V2DeepSeek2025-11685BOpen
Kimi K2 ThinkingMoonshot2025-111.0T *Open
P1-235B-A22BShanghai AI Lab2025-11235B *Open
Ling-1TAnt Group2025-101.0TOpen
MiniMax-M2MiniMax2025-10229BOpen
Tongyi DeepResearchAlibaba2025-1031BOpen
AgentFounder-30BAlibaba2025-0930BOpen
GLM-4.6Z.ai (Zhipu AI)2025-09357B *Open
LongCat-FlashMeituan Inc2025-09560B *Open
MindLink-72BKunlun Inc.2025-0872BOpen
GLM-4.5Z.ai (Zhipu AI)2025-07355BOpen
Kimi K2Moonshot2025-071.0TOpen
Qwen3-235B-A22B (Jul 2025)Alibaba2025-07235B *Open
Qwen3-235B-A22B-Thinking (Jul 2025)Alibaba2025-07235B *Open
Qwen3-Coder-480B-A35BAlibaba2025-07480BOpen
Qwen3 EmbeddingAlibaba2025-068BOpen
DeepSeek-R1 (May 2025)DeepSeek2025-05671BOpen
Qwen3-235B-A22BAlibaba2025-04235B *Open
DeepSeek-V3 (Mar 2025)DeepSeek2025-03671BOpen
QwQ-32BAlibaba2025-0333B *Open
DeepSeek-R1DeepSeek2025-01671BOpen
DeepSeek-V3DeepSeek2024-12671BOpen
Hunyuan-LargeTencent2024-11389BOpen
DeepSeek-V2.5DeepSeek2024-09236BOpen
Qwen2.5 Instruct (72B)Alibaba2024-0973BOpen
Qwen2.5-32BAlibaba2024-0933BOpen
Qwen2.5-72BAlibaba2024-0973BOpen
Telechat2-115BChina Telecom2024-09115BOpen
DeepSeek-Coder-V2 236BDeepSeek2024-06236BOpen

An asterisk marks a parameter figure the source does not consider traceable to the people who built the model. That judgement is the source’s and it has false negatives: Kimi K3 is marked speculative at 2.8T while Moonshot’s own model card states the figure. Why that distinction carries the whole exhibit.

Counts are of models Epoch AI records as notable, not of every model released. The most recent month is always incomplete, because a model enters the dataset when it is reviewed rather than when it ships.

Epoch AI, 'Data on AI Models'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/data/ai-models-documentation' Used under CC BY 4.0.

How this museum is made